Body:
Hello everyone, I encountered an issue while using the Base KNIME Hub and would like to ask for advice:
- Our company purchased Base KNIME Hub, and according to the official documentation:
Using a Basic license automatically provides 4 vCores for execution. Extra execution vCores can be purchased, but in such cases, the available vCores for the system must be calculated accordingly.
- After logging into the KNIME Hub admin console, it correctly shows 4 vCores, consistent with our license.
- However, in the Execution resources section, it only shows 2 vCores, and I am unable to click the “+” to increase or manually input a number.
Could anyone advise if this is a system limitation or a configuration issue? I have attached detailed screenshots for reference.
Thank you in advance for your help!
Hi @DanielHua ,
as far as I can conclude from your screenshot your already have the default “Initial Execution Context”. Since that already got two vCPUs allocated you cannot allocate more than totally available.
Please let me know if I interpretated that incorrectly.
Best, Mike
CTO @ DataNautics GmbH - Your KNIME-Experts
Contact: info@datanautics.gmbh // datanautics.gmbh // +49(0)170-325 713 9 // Linkedin
Daten Automatisierung für Finanz-, Produktion-, IT- und Marketing-Prozesse mit KNIME
1 Like
Hi Mike,
Thanks a lot for your reply!
Yes, the Initial Execution Context currently uses 2 vCores.
However, what confuses me is that our KNIME Hub Base license includes 4 vCores, and the Hub Admin page also confirms a total of 4 vCores available.
According to the official KNIME documentation (Business Hub Installation Guide):
Using a Basic license automatically provides 4 vCores for execution. Extra execution vCores can be purchased, but in such cases, the available vCores for the system must be calculated accordingly.
So I was expecting that I could allocate up to 4 vCores to the execution context, but it seems limited to 2 only.
Could this be due to one of the following?
- The Executor (container or VM) itself is configured with only 2 vCores?
- Or the Base license allows a total of 4 vCores across the system, but limits each Execution Context to 2 vCores?
If needed, I can also share our Executor configuration details for verification.
Appreciate your clarification!
I’ve attached the screenshot of the Executor resource allocation for reference.
Best regards,
Daniel
Hi Daniel,
we faced an issue lately where none of the >16 vCPUs got recognized et all but never the case of being unable to assign more than just two. The license type might indeep be a potential cause.
The docs I interprete in such a way that per executor you can assign a max. of 4vCPUs of a total of the recommended 16, minus 10 for Hub Core.
The docs, a few pixels up point towards the allocation “Total Capacity” (16). Substract 10 vCPUs for the Hub Core leaving you with 6 out of which 4 are provisioned within the “Initial Execution Context” leaving a spare of 2 vCPUs. It appears you are creating a second execution context, aren’t you? If correct, the Hub only allows to allocate the remaining two.
The CPU and Memory requirement amounts shown in the following chapters of this document refer to the “Total Capacity” of the nodes. Hence not all of these resources are allocatable for Kubernetes workloads, some needs to be available or reserved for system-related-services and should not be allocated, or reserved for Kubernetes related workloads. By default usually only some Memory is reserved from the total capacity, therefore if there are no reservation for CPU vCores, it is advised to either reserve or leave some vCores un-reserved by Kubernetes workloads. The reserved Memory for system-related-services is usually 100 MB by default for Kurl installations.
Can you try to modify the Initial Execution Context and reduce the vCores to three and then try to create a new executor with three vCPIs?
Cheers
Mike
1 Like
Hi Daniel,
vCores can be assigned not only to Execution Contexts, but to teams as well.
Can you please check your current team’s resources to see if they have 2 vCores/core tokens already claimed?
Example:
If yes, that would mean those 2 cores are reserved for the team and cannot be given to the Execution Context - you will need to reduce the team core tokens to 0 to free them up.
2 Likes
Very valid point too but weren’t these appear to be highliged as blue? Maybe there is some inconsistency here …
1 Like
Hi Mike,
Thank you very much for your detailed explanation — it really helped me understand the resource allocation mechanism in KNIME Hub more clearly.
The current Execution Context shows 2 vCores allocated. The “+” button is greyed out, and manually entering the number 4 has no effect. I will follow your suggestion and try changing it to 3 to see if it works.
I’ll test this configuration and check if the allocation refreshes correctly.
Thanks again for your continued support — I’ll report back once I have the results.
Best regards,
Daniel
1 Like
